Physical brain dynamics: Activity, structure, and function via neural fields and eigenmodes
By viewing the brain as a multiscale physical system one can model and analyze its activity and resulting measurements quantitatively. Physiologically based neural field theory (NFT) permits tractable prediction and analysis from sub-mm scales to the whole brain. The results reproduce multiple experimental outcomes in normal and abnormal states, ranging from spontaneous activity to stimulus responses, evolution of sleep-wake cycles, brain plasticity, and epileptic seizures; Measurement techniques covered include electroencephalography, functional MRI, and anatomical connectivity tracing. Normal brain operation is found to be near-critical, with seizures setting in via Hopf bifurcations to limit cycle dynamics, and driven brain activity exhibiting a rich range of nonlinear dynamics. NFT brain eigenmodes exhibit structure that is closely related to spherical harmonics and their dynamics can be used to simplify analysis of neural activity. NFT then allows brain connectivity to be inferred noninvasively from functional activity correlations via eigenmode analysis, including connectivity that cannot be observed directly. The outcomes illustrate the power of physically based modeling to predict and unify multiple observations across scales. They open the way to a host of new applications and physically based analysis methods.